Sharper Image LM521 Atomic Projection Clock

Transmitter Repair and Deconstruction
Atomic Clock
My clock is very similar in appearance but is Sharper Image badged.
It is the same as the La Crosse WT-5721.

I received this clock as a gift in 2002, and in 2015, I was noticing weaker output from the transmitter and then eventually the outside temperature was not updating.  The WWV radio update (atomic operation) and the projection of time and temperature are our favorite features.

The transmitter is sealed, and I started by sawing with a thin serrated blade along the case seam.  There were no inner attachment points, and the case opened up at that point.  The front and back view of the board are shown below.

Front of circuit board
The front of the board.  One can see the two battery holders on the botom.

Back of board
Back view of the board.  You can see the coil antenna as the top right curved trace.

Referring to the back of the board, it is easy to see that the large trace on the bottom of the image and the square at the top is the ground reference plane.  The processor is clearly encapsulated in the black blob, and the RF oscillator/transmitter is at the top.  This latter module is controlled probably by the crystal in the can on the front of the board.  One can also see an antenna on the board traces at the top-right coupled to by a variable capacitor. It appears (since there is a crystal) that the capacitor adjust the antenna tuning for maximum output (and should not affect the frequency).  The actual temperature sensor appears to be a small part in the front middle of the board that sits in a hole in the front of the small transmitter box.

There are fortunately only two traces that cross into the square ground plane at the top, and one of which is power, the other is easily determined to be serial output.  These signals are easily accessed by the three adjacent round holes in the middle of the board.  These are power (left on back), ground, and serial output (right on back).  

Putting the signal on the serial output line on the scope shows that the following data protocol (at room temperature).
  • There are three groups of data transmissions.  Each last about 85-87 msec with a 24 msec pause between them.
  • Zooming into each group, it is clear that all the 'low' periods are the same duration, and there are two 'high' durations.  This is similar to the Sony SIRCS protocol, and so we can characterize a '1' bit is a long high and a low, while the '0' bit is a short high and a low.
  • The timing of the '1' is  1.36 msec high, and 920 usec low.  The '0' is 560 usec high and 920 usec low.  For 50% duty cycle data, the average bit time is 1.9 msec, so there are about 46 bits per group.
  • When the board is initially powered up, the chip transmits a packet every six seconds.  After about four minutes, the output rate drops to about one per minute.
Cleaning the whole assembly (especially the battery contacts) and then reassembling showed that the transmitter worked fine.  As a mod, I soldered a few inches of thin wire-wrap wire to the top terminal on the back as an external antenna.  This mod appears to make the signal strength much better.  I suspect the frequency is 433 Mhz, and the length of the quarter-wave antenna is 6.5", but I did not make my antenna quite that long (about 5").

April 2024
The clock stopped working one day and I found out after some poking around that one of the battery contacts (+ supply) was cracked and needed some care.  After desoldering, cleaning and reinstalling it worked again.  I was aided by my Flipper device during this repair.

After power up and scanning with the "Weather Station" app on the Flipper, I get an entry identified as being at 433.92 MHz and AM Modulated.

Flipper scan
Transmitter shows up at 433.92 MHz

Data transmitted has the temperature data embedded in it.

Clicking on this entry allows me to view the data contents.  I presume the "LaCrosse_TX 44b" is some kind of data format (the La Crosse model number of the transmitter is TX6U).  It appears that the transmitter uses a random serial number (0x3E here) on startup, and the receiver will start using this one after it is power cycled.  That explains why I have to power cycle the clock when I change the batteries on the transmitter.  Presumably this allows the use of 256 different transmitters in the same local area.  Each one is powered up and registered in sequence by the receiver.

The data packet display above is probably some kind of header and contains the temperature data.  The Flipper decoded this information and shows it as 26.4C. It also resets a clock that shows the amount of time since the previous packet was sent.

Main Home Automation Page

Back Home